U.S. patent number 5,999,802 [Application Number 08/865,586] was granted by the patent office on 1999-12-07 for direct conversion tuner.
This patent grant is currently assigned to RCA Thomson Licensing Corporation. Invention is credited to Felix Aschwanden.
United States Patent |
5,999,802 |
Aschwanden |
December 7, 1999 |
Direct conversion tuner
Abstract
A direct conversion tuner for tuning either analog or digital
television signals includes a first and second channels, each
having first and second mixers and an intervening filter stage,
coupled between an RF input and an output combining unit. The first
mixers receive respective first local oscillator signals which have
the same frequency but a quadrature phase relationship. The
frequency of the first local oscillator signals is controlled
according to the selected channel so that it is located within the
spectrum of the respective RF signal. The second mixers receive
respective second local oscillator signals which have the same
frequency but a quadrature phase relationship. The frequency of the
second local oscillator signal is located above the passband of the
filter stages. A digital gain and phase equalization network is
included in one of the channels for adjusting the relative gain and
phase shift of the two channels and is automatically controlled by
a microcomputer in response to signals sampled at respective points
within the first and second channels to reduce the relative gain
and phase shift. As a result, near perfect cancellation of unwanted
components occurs in the output combining unit.
Inventors: |
Aschwanden; Felix (Thalwil,
CH) |
Assignee: |
RCA Thomson Licensing
Corporation (Princeton, NJ)
|
Family
ID: |
26303009 |
Appl.
No.: |
08/865,586 |
Filed: |
May 29, 1997 |
PCT
Filed: |
June 06, 1994 |
PCT No.: |
PCT/IB94/00138 |
371
Date: |
November 22, 1995 |
102(e)
Date: |
November 22, 1995 |
PCT
Pub. No.: |
WO94/29948 |
PCT
Pub. Date: |
December 22, 1994 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
553264 |
|
|
|
|
|
Foreign Application Priority Data
|
|
|
|
|
Jun 4, 1993 [GB] |
|
|
9311610 |
Apr 26, 1994 [GB] |
|
|
9408211 |
|
Current U.S.
Class: |
455/196.1;
455/226.1; 455/242.2; 455/324; 348/731 |
Current CPC
Class: |
H03D
3/009 (20130101); H03D 3/008 (20130101); H03D
7/166 (20130101); H03D 2200/005 (20130101) |
Current International
Class: |
H03D
3/00 (20060101); H03D 7/00 (20060101); H03D
7/16 (20060101); H04B 001/26 () |
Field of
Search: |
;455/67.1,67.4,150.1,186.1,196.1,242.1,242.2,243.1,265,246,303,304,306,307,316
;375/324,337,344,349,350 ;331/25 ;488/226.1 |
References Cited
[Referenced By]
U.S. Patent Documents
Foreign Patent Documents
|
|
|
|
|
|
|
0343273 |
|
Nov 1989 |
|
EP |
|
0496621 |
|
Jul 1992 |
|
EP |
|
Other References
Proceedings of the IRE, "A Third Method of Generation and Detection
of Single-Sideband Signals", D.K. Weaver, Jr. pp. 1703-1705. .
1993 IEEE, "Digital TV Receiver with a Low IF", Dietmar Ehrhardt et
al, pp. 331-339..
|
Primary Examiner: Eisenzopf; Reinhard J.
Assistant Examiner: Kincaid; Lester G.
Attorney, Agent or Firm: Tripoli; Joseph S. Emanuel; Peter
M. Akiyama; Kuniyuki
Parent Case Text
This is a continuation of application Ser. No 08/553,264, filed
Nov. 22, 1995 now abandoned, which is a 371 of PCT/IB94/00138,
filed Jun. 6, 1994.
Claims
I claim:
1. In a receiver, tuning apparatus for tuning a selected one of a
plurality of RF signals received at an RF input to produce an
output signal at an output, comprising:
first and second channels each having an input and an output and,
in the order named, a first mixer stage, a filter stage and a
second mixer stage coupled between said input and output;
said inputs of said first and second channels being coupled to said
RF input;
a summing unit having first and second inputs and an output, said
outputs of said first and second channels being coupled to
respective inputs of said summing unit, and said output of said
summing unit being coupled to the output of said tuning
apparatus;
means for providing first local oscillator signals of the same
frequency but of quadrature phases to respective ones of said first
mixer stages; said frequency of said first local oscillator signals
being located within the frequency spectrum of said selected RF
signal;
means for providing second local oscillator signals of the same
frequency but of quadrature phases to respective ones of said
second mixer stages; said frequency of said second local oscillator
signals being located above the passbands of the respective said
filter stages;
means for monitoring first and second signals produced at
respective points within said first and second channels;
means for adjusting the relative gain and the phase shift of said
first and second channels prior to summing of said outputs of said
first and second channels in said summing unit;
said gain and phase shift adjusting means comprising respective
individual units and being included in the same one of said first
and second channels;
means for automatically controlling said gain and phase shift
adjusting means to reduce the differences between the relative
amplitudes and the phases of said first and second channels in
response to the relative amplitudes and phases of said first and
second signals, wherein
said gain adjusting means is coupled in cascade with an additional
summing network between said second mixer and said output of said
one channel; and
said phase shift adjusting means comprises an additional mixer and
an additional gain adjusting means coupled in cascade between said
filter stage and said output of said one channel; said additional
mixer receiving a signal having the same frequency as the frequency
of said second local oscillator signal of said one channel but of
quadrature phase.
2. In a receiver, tuning apparatus for tuning a selected one of a
plurality of RF signals received at an RF input to produce an
output signal at an output, comprising:
first and second channels each having an input and an output and,
in the order named, a first mixer stage, a filter stage and a
second mixer stage coupled between said input and output;
said inputs of said first and second channels being coupled to said
RF input;
a summing unit having first and second inputs and an output, said
outputs of said first and second channels being coupled to
respective inputs of said summing unit, and said output of said
summing unit being coupled to the output of said tuning
apparatus;
means for providing first local oscillator signals of the same
frequency but of quadrature phases to respective ones of said first
mixer stages; said frequency of said first local oscillator signals
being located within the frequency spectrum of said selected RF
signal;
means for providing second local oscillator signals of the same
frequency but of quadrature phases to respective ones of said
second mixer stages; said frequency of said second local oscillator
signals being located above the passbands of the respective said
filter stages;
means for monitoring first and second signals produced at
respective points within said first and second channels;
said means for monitoring said first and second signals including
first and second analog-to-digital converters respectively;
said first and second digital-to-analog converters being coupled
between respective ones of said filters and said second mixer
stages of said first and second channels;
said second mixer stages of respective ones of said first and
second channels being digital mixers;
means for adjusting the relative gain and the phase shift of said
first and second channels prior to summing of said outputs of said
first and second channels in said summing unit;
means for automatically controlling said gain and phase shift
adjusting means to reduce the differences between the relative
amplitudes and the phases of said first and second channels in
response to the relative amplitudes and phases of said first and
second signals;
said means for automatically controlling said gain and phase shift
adjusting means including a microcontroller operating under program
control and being coupled to said analog-to-digital converter,
wherein
said automatic controlling means includes means for generating a
reference signal and inserting it into each of said first and
second channels after respective ones of said first and second
mixers.
3. Tuning apparatus for tuning to a selected one of a plurality of
RF signals received at an RF input to produce an output signal at
an output, comprising:
first and second channels each having an input and an output and,
in the order named, a first mixer stage, a filter stage and a
second mixer stage coupled between said input and output;
said inputs of said first and second channels being coupled to said
RF input;
a summing unit having first and second inputs and an output, said
outputs of said first and second channels being coupled to
respective inputs of said summing unit, and said output of said
summing unit being coupled to the output of said tuning
apparatus;
means for providing first local oscillator signals of the same
frequency but of quadrature phases to respective ones of said first
mixer stages; said frequency of said first local oscillator signals
being located within the frequency spectrum of said selected RF
signal;
means for providing second local oscillator signals of the same
frequency but of quadrature phases to respective ones of said
second mixer stages; said frequency of said second local oscillator
signals being located above the passbands of the respective said
filter stages;
means for monitoring first and second signals produced at
respective points within said first and second channels;
means for adjusting the relative gain and the phase shift of said
first and second channels prior to summing of said outputs of said
first and second channels in said summing unit; and
control means coupled to said monitoring and adjusting means for
inserting a reference signal into both channels for automatically
controlling said gain and phase shift adjusting means to reduce the
differences between the relative amplitudes and the phases of said
first and second channels in response to the relative amplitudes
and phases of said first and second signals.
4. The tuning apparatus recited in claim 3, wherein:
said reference signal has a plurality of frequencies; and
said control means is responsive to said first and second signals
as affected by said reference signal at each of said pluralities of
said frequencies.
5. The tuning apparatus recited in claim 3, wherein:
said reference signal has a plurality of frequencies and said
automatic controlling means is responsive to said first and second
signals as affected by said reference signal at each one of said
pluralities of said reference signal.
6. The tuning apparatus recited in claim 3, wherein:
said gain adjusting means is coupled in cascade with an additional
summing unit between said second mixer stage and said output of
said one channel;
said phase shift adjusting means comprises an additional mixer
stage and an additional gain adjusting means coupled in cascade
between said filter stage and said output of said one channel;
and
said additional mixer stage receives a signal having the same
frequency as the frequency of said second local oscillator signal
of said one channel but of quadrature phase.
Description
FIELD OF THE INVENTION
The invention concerns a so called "direct conversion tuner" which
is especially useful in a television receiver.
BACKGROUND OF THE INVENTION
An early type of tuner known as a "tuned radio frequency tuner"
(TRF) included several radio frequency (RF) amplifiers which were
all tuned to the frequency of the RF signal of desired transmission
channel followed directly by a detection section, without an
intervening mixer employed in later tuners. Such a tuner could
provide relatively distortion free performance due to the absence
of a mixer. However, TRF tuners tended to be large in size and
subject to stability and gain control problems due to the number RF
amplifiers which were needed. Moreover, TRF tuners did not provide
a consistent or adequate degree of signal selectivity.
The type of tuner which is primarily used today is known as a
"heterodyne" or "superheterodyne" tuner. In its simplest form,
known as a "single-conversion" tuner, it comprises a tunable RF
amplifier followed by a frequency conversion stage, including a
mixer and a local oscillator. The frequency conversion stage
produces an intermediate frequency (IF) signal which corresponds to
the received RF signal but has a lower frequency. The IF signal is
filtered by an IF filter section and the resultant signal is
coupled to a detection section. The combination of conversion stage
and the following IF filter section provides a significantly better
selectivity characteristic than a TRF tuner. The frequency of the
local oscillator signal is offset (usually higher) from the
frequency of the desired RF signal by the desired frequency of the
IF signal. In a television receiver, the local oscillator signal is
controlled so that it places the frequency of the IF picture
carrier corresponding to the RF picture carrier at a nominal
frequency, for example, at 45.75 MHz in the United States and 38.9
MHz in Europe.
A single conversion tuner can be made quite small and relatively
inexpensive. However, it produces unwanted intermodulation and
cross-modulation products due to third and higher order components
of the signal transfer characteristics of the mixer included in the
frequency conversion stage. Various unwanted conversion products,
known in the tuner fields as "image", "one-half IF" and "IF beats",
continue to be a problem. The IF filter is designed to minimize
unwanted conversion products and also to provide rejection of
responses due to adjacent channels (selectivity). Thus, the
selection of the IF frequency is a compromise. As a result the
rejection of unwanted conversion products and selectivity of the
tuner may not be adequate.
The deficiencies of a single conversion tuner have become
especially troublesome due to the increasing number of "contiguous"
channels now available in large cable television systems. With the
advent of digital television transmission systems, such as for high
definition television (HDTV), the problem becomes still more
difficult because these systems make full use of the available
channel spectrum and only a small guard band of a few hundred
kilohertz (KHz) exists between channels. In addition, the overall
frequency response of a tuner for tuning digital television signals
must be flat to the edges of the channel, but nevertheless, have a
very steep "roll-off" (attenuation) at the edges for adequate
adjacent channel rejection. This makes the design of an appropriate
IF filter more complicated since no Nyquist slopes and sound traps,
which tend to ease IF filter design, can be used in digital
systems. In addition, it is contemplated that both analog and
digital television signals will be transmitted during a transition
period. In that case, even more adjacent channel selectivity will
be required for good reception of the digital signals because
digital television signals will be transmitted with much less power
than analog television signals.
The "double-conversion" variation of the superheterodyne tuner was
developed to overcome the shortcomings of the single-conversion
tuner. In this type of tuner, a first conversion stage is followed
by a first IF filter section, a second conversion stage, and a
second IF filter stage. The first IF section has a very high
frequency range, typically in the order of 620 MHz. The second IF
section has a much lower frequency range, typically the same as
that of the only IF filter section of a single conversion tuner.
The second IF section is followed by a detection section.
The very high frequency of the first IF filter section places the
RF signals corresponding to unwanted conversion responses such as
the "image" response at frequencies readily rejected by tunable RF
stages which precede the first conversion stage. The low frequency
second IF provides the required adjacent channel selectivity needed
for modern television reception. Unfortunately, an
double-conversion tuner system requires additional RF and IF
circuitry compared to a single conversion tuner, and much of the
additional circuitry must function at relatively high frequencies
requiring extensive shielding. As a result a double conversion
tuner is relatively large in size and expensive.
Another type of tuner known as a "direct conversion" tuner has
improved unwanted conversion product rejection and selectivity
properties with respect the TRF and heterodyne types of tuners. A
direct conversion tuner operates in accordance with a third tuning
method in which the frequency of a local oscillator signal of a
first frequency conversion stage is set in the middle of the
frequency band of the desired channel. The product of the first
conversion stage is at relatively low frequency. There are no image
responses because the frequency of the first conversion stage is
located with the spectrum of the desired RF signal. In addition,
the very low frequency range of the signal produced at the output
of the first conversion stage makes it possible to readily provide
a filter which can reject adjacent channel signals.
Unfortunately, because the first local oscillator signal is
centered in the frequency band of the desired channel, both the
upper and lower side band of the desired channel will be converted
to the frequency range of the first IF signal so that the lower
side band (LSB) is in effect folded over onto the upper side band
(USB) in the spectrum of the first IF signal. Since the LSB and USB
occupy the same frequency range, the LSB and the USB must again be
separated before detection. To accomplish this, a direct conversion
tuner is arranged as is shown in FIG. 1.
Basically, the direct conversion tuner contains two channels, each
with two conversion stages. The received RF signal is coupled to
each of two mixers M1A and M1B via a tuned RF amplifier which
provides gain and some selectivity. Desirably, the gain of the RF
amplifier is automatically controlled in response to an automatic
gain control (AGC) signal (not shown). The local oscillator signal
generated by a first local oscillator LO1 is tuned to the center
frequency too of the frequency band of the desired channel between
the lower sided band (LSB) and the upper side band (USB), as is
shown in FIG. 2a. The first local oscillator signal is split by a
phase shifting circuit PS1 into quadrature components that are used
to drive mixers M1A and M1B. The respective IF output signals of
mixers M1A and M1B are filtered by two low pass filters LPF A and
LPF B. Low pass filters LPF A and LPF B provide the necessary
selectivity to reject the responses from the adjacent channels and
higher order products of mixers M1A and M1B.
Each of the output signals of mixers M1A and M1B includes both a
lower side band portion and an upper side band portion
corresponding to the LSB and USB portions of the received RF
signal. However, as earlier indicated, the LSB portion is folded
over so that it is superimposed on the USB portion and occupies the
same frequency range, as is shown in FIG. 2b. The output signal of
low pass filters LPF A and LPF A are coupled to respective ones of
a second pair of mixers M2A and M2B. Mixers M2A and M2B are driven
by respective ones of a second pair of quadrature local oscillator
signals generated by a second local oscillator LO2 and a second
phase shifting circuit PS2. Each of the second local oscillator
signals has a frequency .omega..sub.N located above the cutoff
frequency of the low pass filters LPF A and LPF B filters to
fulfill the Nyquist criteria. The output signals of mixers M2A and
M2B are added in a summer unit SU to produce an output signal which
has a spectrum which includes separated lower and upper side band
portions, as shown if FIG. 2c. This output signal is coupled to a
demodulator (not shown) which demodulates it, and the demodulated
resultant is coupled to further signal processing sections.
The operation of the direct conversion tuner shown in FIG. 1 can be
mathematically understood by considering a very simple case in
which the received RF signal is assumed to include a sinusoidal
upper side band component of sin (.omega..sub.0 +.omega..sub.1) and
a sinusoidal lower side band component of sin (.omega..sub.0
-.omega..sub.2), as is indicated in FIG. 3a. It is also assumed
that the gains and phase shifts of the two channels are identical.
The phases of the signal components produced at various point of
the direct conversion tuner are indicated by the vector arrows in
FIG. 1. Further, the coefficients of the various mathematically
factors corresponding to signal components have been normalized in
the following description.
The quadrature first local oscillator signals applied to first
mixers M1A and M1B are expressed as sin .omega..sub.0 and cos
.omega..sub.0, respectively; and the quadrature second local
oscillator signals applied to second mixers M2A and M2B are
expressed as sin .omega..sub.N and cos .omega..sub.N, respectively.
The following signal is produced at the output of low pass filter
LPF A:
The following signal is produced at the output of low pass filter
LPF B:
The spectra at the outputs of low pass filters LPF A and LPF B are
shown in FIG. 3b.
The result of the second mixing operation by mixer M2A produces the
following output signal:
The result of the second mixing operation by mixer M2B produces the
following output signal:
The addition of the two output signals of mixers M2A and M2B by
summer SU results in the following signal:
The spectrum at the output of summer SU is shown in FIG. 3c.
The operation of the direct conversion tuner depends on the
cancellation of unwanted components developed in the two channels
(compare the output signals of mixers M2A and M2B indicated above
including the terms sin (.omega..sub.N -.omega..sub.1) and sin
(.omega..sub.N +.omega..sub.2)). As was stated above, the
description of the operation of the direct conversion tuner so far
provided assumes that the gains and phase shifts of corresponding
elements of the two channels are identical, resulting in perfect
cancellation of the unwanted components after the addition of the
output signals of the two channels by summer SU. However, in
practice, gain and phase characteristics of the two channels are
unequal and change with temperature and time. The gain and phase
characteristics affect the phase and magnitude of the vectors shown
in FIG. 1. As a result, perfect cancellation of the unwanted
components no longer occurs causing the generation of unwanted
spurious components in the output signal produced by summer SU and
the reduction of the quality of the demodulated signal. This is
especially the case when the received RF signal is relatively
complex, such as a television signal, and does not simply contain a
lower and an upper sinusoidal component as assumed in the above
description.
The generation of unwanted spurious components when a television
signal is tuned by a direct conversion tuner of the type shown in
FIG. 1 is illustrated in FIGS. 4a, 4b and 4c. FIG. 4a shows the
spectrum of a television signal of a single channel. It includes a
picture carrier (PIX), a color subcarrier (SC) and a sound carrier
(SOUND). The frequency, .omega..sub.0, of the first local
oscillator signal is located approximately midway between the
picture carrier and the sound carrier. FIG. 4b shows the spectrum
of the signal resulting from the first mixing operation. FIG. 4c
shows the spectrum of the output signal of summer SU. For each
desired component of the output signal of summer SU to the right of
the frequency, w.sub.N, of the second local oscillator signal, an
undesired "companion" to the left exists; and for each desired
component of output signal of summer SU to the left of frequency
.omega..sub.N an undesired "companion" to the right exists. For
instance, a "companion" of the picture carrier is present to the
right of .omega..sub.N between the color subcarrier and the sound
carrier. The presence of the unwanted "companions" causes annoying
beat patterns in the demodulated video signal and may also
adversely affect the demodulated sound signal. Such unwanted
components should desirably be suppressed in the order of 45 to 50
dB for optimum performance of the television receiver. This means
that the gain and phase errors should desirably be kept less than
0.05 dB. and 0.5 degrees, respectively, for optimum performance of
the television receiver. Such performance standards cannot be
obtained and maintained with manual adjustments.
SUMMARY OF THE INVENTION
The present invention concerns an arrangement for automatically
reducing the gain and phase difference errors of the two channels
of a direct conversion tuner in order to reduce the generation of
unwanted components in the output signal. In accordance with an
aspect of the invention, a multiple frequency reference signal is
inserted as a test signal at respective insertion points of the two
channels and signals produced at respective measurement points of
the two channels are compared as to amplitude and phase to produce
so called "ripple" gain and phase difference responses related to
the gain and phase shift differences between two channels, for
example, due to the IF low pass filters. In accordance with another
aspect of the invention, the signals developed at the measurement
points in response to the RF signal for the selected channel are
also compared as to amplitude and phase to produce so called "DC"
gain and phase difference values related to gain and phase shift
differences of one or more of the conversion stages. The "ripple"
gain and phase responses and the "DC" gain and phase difference
values are used to control a gain and phase correction network so
that the amplitude and phase differences between the measured
signals of the two channels are reduced. Preferably, the
arrangement is implemented in a digital embodiment in which
analog-to digital converters are included in respective ones of the
channels after the first conversion stages and the gain and phase
correction units comprise digital filters after the second
conversion stages. The operation of automatic gain and phase
equalization arrangement can be automatically initiated each time
the television receiver is switched on or a new channel is
selected.
These and other aspects of the invention will be described with
respect to the accompanying Drawings.
BRIEF DESCRIPTION OF THE DRAWING
In the Drawing, FIGS. 1, 2a-2c, 3a-3c and 4a-4c concern the
background of the invention and have previously been described.
Briefly:
FIG. 1 is a block diagram of a direct conversion tuner as is known
in the prior art;
FIGS. 2a, 2b and 2c are graphical representations of the spectra of
signals at various points of the direct conversion tuner shown in
FIG. 1;
FIGS. 3a, 3b and 3c are graphical representations of the spectra of
signals at various points of the direct conversion tuner shown in
FIG. 1 assuming that the input signal consists of two sinusoidal
components; and
FIGS. 4a, 4b and 4c are graphical representations of the spectra of
signals at various points of the direct conversion tuner shown in
FIG. 1 when the input signal comprises an analog television
signal.
The remaining Figures of the Drawings concern the embodiment of the
invention. Briefly:
FIG. 5 is a block diagram of a direct conversion tuner including an
automatic gain and phase equalization arrangement constructed in
accordance with an aspect of the invention;
FIG. 6 is a vector diagram useful in understanding how gain and
phase errors are measured by the automatic gain and phase
equalization arrangement;
FIGS. 7a and 7b, 8a and 8b, and 9a and 9b are graphical
representations of spectra of phase and gain errors useful in
understanding how phase and gain correction information used by the
automatic gain and phase equalization arrangement is obtained;
FIG. 10 is a vector diagram also useful in understanding how phase
and gain correction information used by the automatic gain and
phase equalization arrangement is obtained;
FIGS. 11a and 11b are graphical representations of spectra of phase
and gain correction responses needed for phase and gain
equalization which result from the operations illustrated in FIGS.
7a and 7b, 8a and 8b, 9a and 9b and 10;
FIG. 12 is a flow chart indicating the overall operation of the
automatic gain and phase equalization arrangement.
DETAILED DESCRIPTION OF THE DRAWING
The direct conversion tuner shown in FIG. 5 is generally similar to
the one shown in FIG. 1, but includes additional elements which
comprise an automatic gain and phase equalization arrangement
constructed in accordance with an aspect of the invention and
certain related elements. Those elements of the direct conversion
tuner shown in FIG. 5 which have the same or similar functions as
corresponding elements of the direct conversion tuner shown in FIG.
1 are identified by the same or similar reference designations and
will not be described in details again.
The automatic gain and phase equalization arrangement comprises a
gain and phase correction network including a first gain control
unit labeled GC and a second gain control unit which is actually
used for phase correction and is therefore labeled PC. The gain and
phase correction network is inserted in channel A between LPF A and
output summer SU2. Gain correction unit GC and phase correction
unit PC comprise respective programmable digital filters which are
controlled by a microcomputer MC to adjust the gain and phase
characteristics of channel A so that the gain of phase
characteristics of the channels A and B are substantially
identical. Gain correction unit GC and phase correction unit PC
may, for example comprise, finite impulse response (FIR) filters.
Microcomputer MC samples the signals developed at points A and B
just before second mixers M2A and M2B in response to a test signal
and determines the relative amplitudes and phases of the sampled
signals to develop filter coefficient control signals for gain
correction unit GC and phase correction unit PC. For this purpose,
microcomputer MC generates a reference signal labeled
.omega..sub.REF which is inserted as a test signal just after first
mixers M1A and M1B, for example, via resistors RA and RB.
Microcomputer MC also samples the signals developed at points A and
B in response to the received RF signal for the selected channel
and determines the relative amplitudes and phases of the sampled
signals the to develop the filter coefficient control signals. The
generation of the filter coefficient control signals will be
described in detail below. Microcomputer may comprise the same
microcomputer which is used to control other functions of the
television receiver.
In the embodiment of invention shown in FIG. 5, the gain and phase
correction is accomplished by the addition of two signals which are
in quadrature phase relationship as will be described below in
greater detail with respect to the vector diagram shown in FIG. 10.
For this purpose gain correction unit GC and phase correction unit
PC (which is actually another gain control unit) are included in
separate paths for which the respective signals are in quadrature
phase relationship. More specifically, gain correction unit GC
receives a signal developed at a point C at the output of second
mixer M2A while phase correction unit PC receives a signal
developed at a point F at the output of an additional mixer M3. The
same local oscillator signal which is coupled to second mixer M2B
of channel B is coupled to additional mixer M3 so that oscillator
signals coupled to mixers M2A and M3 have a quadrature phase
relationship (sin .omega..sub.N and cos .omega..sub.N,
respectively). As a result, the signals developed at points C and F
have a quadrature phase relationship. The output signal of gain
control unit GC developed at a point C' is added to the output
signal of phase correction unit PC developed at a point F' by an
additional summer unit SU2. The signal produced at a point G at the
output of additional summer unit SU2 is combined with to the signal
produced at a point D at the output of a mixer M2B by a summer unit
SU1 corresponding to summer unit SU of the direct conversion tuner
shown in FIG. 1. The signal developed at a point E at the output of
summer SU1 is coupled to a demodulator section (not shown).
Since the gain and phase equalization apparatus is implemented in
digital form in the direct conversion tuner shown in FIG. 5,
analog-to-digital converters ADC A and ADC B have been added to
respective ones of channels A and B prior to the second conversion
stages. The ADC A and ADC B receive amplified versions of the
output signals of LPF A and LPF B from respective ones of
amplifiers AMP A and AMP B and provide respective digital versions
of the output signals of LPF A and LPF B. The outputs of amplifiers
AMP A and AMP B are capacitively coupled to the inputs of the ADC A
and ADC B via respective ones of capacitors CA and CB to avoid DC
drift problems. The amplifiers may be omitted when the outputs
signals of LPF A and LPF B have sufficient amplitudes to permit
reliable analog-to-digital conversion. In the direct conversion
tuner shown in FIG. 5, second mixers M2A and M2B comprise digital
multipliers instead of analog mixers and summers SU1 and SU2 are
digital adders.
Before describing the gain and phase equalization operation in
detail it is helpful to describe the nature of the gain and phase
errors. The gain and phase errors can be divided into "ripple" and
"DC" errors. The "ripple" error is caused by the differences
between the gain and phase responses of low pass filters LPF A and
LPF B. It varies as a function of the IF frequency, but is constant
for all received television channels. The "DC" error is caused by
quadrature errors and gain differences between first mixers M1A and
M2B. It is constant for a selected channel but varies with the
frequency of the first local oscillator. Second mixers M2A and M2B
do not introduce additional errors if they are implemented in
digital form as is the case in the direct conversion tuner shown in
FIG. 5.
There is an additional error which can be referred to as an
"asymmetry" error which occurs when first mixers M1A and M1B are
implemented as a doubly-balanced mixer. The use of a
doubly-balanced mixer is desirable since it tends to reduce the
coupling of the RF signal and local oscillator signal to the output
of the mixer because of its balanced configuration. However, a
doubly balanced mixer is not perfectly balanced and this results in
slightly different gain and phase characteristics for RF signal
components which have a frequency below the frequency,
.omega..sub.0, of the first local oscillator signal than for RF
signal components which have a frequency above the frequency of the
first local oscillator signal. The "asymmetry" error is small and
can be neglected in most cases. However, it can be corrected, if
desired, as will be described below.
It is possible to measure gain and phase errors with the use of a
"sweep" generator which is connected to the antenna input, and
which "sweeps" the frequency range of a selected channel. This
method is relatively complicated and expensive. Measuring the
ripple and DC errors separately is simpler.
The ripple error can be measured by using a multiple frequency
reference signal, labeled .omega..sub.REF in FIG. 5. The reference
signal has a low frequency range which only needs to be
sufficiently broad to cover the IF frequency range. Experiments
have shown that eight to ten discrete test frequencies are
sufficient to obtain satisfactory results. The reference signal may
be generated by an oscillator (not shown) under the control of
microcomputer MC or directly by microcomputer MC by using a look-up
table. During the ripple error test, the tuner has to be disabled
from responding to the received RF signal. This may be accomplished
by turning off the first local oscillator (LO1) or disabling the RF
stage. The ripple errors can be measured each time the television
is switched on or when a new channel is selected. Once the ripple
errors have been measured, the normal operation of the tuner is
again initiated.
The DC errors of mixers M1A and M1B are local oscillator or
channel-dependent, but are otherwise constant over the frequency
range of each channel. Therefore, the DC errors can be measured
using a single frequency test signal for each selected channel.
However, advantageously, the picture and sound carriers of the
received television signal can be used to measure the DC errors.
Both carriers are known in frequency and have high energy which
makes the measurement reliable. Using both the picture and sound
carriers allows for the correction of the asymmetry errors because,
as shown in FIG. 4a, the frequency of the picture carrier is below
the frequency, .omega..sub.0, of the first local oscillator signal
and the frequency of the sound carrier is above the frequency of
the first local oscillator signal. The carriers are measured at
points A and B at the outputs of ADC A and ADC B during the
occurrence of the broad vertical equalizing pulses when the picture
carrier has its highest energy but is not modulated with video
information. However, because both picture and sound carriers are
present, the carriers need to be separated. The separation can be
accomplished after the measurement by microcomputer MC in
accordance with a software filter program. For example, low pass
and high pass filter responses can be obtained utilizing a
MatLab.TM. program commercially available from MATHWORKS, Inc., of
Massachusetts. The low and high pass filter responses can be used
to separate the picture carrier and sound carrier responsive
portions of the measurement values. Alternatively, microcomputer MC
may comprise a digital signal processor (DSP) unit including
digital filters and a microprocessor may advantageously be used. In
that case, the digital filters can be used to separate the picture
and sound carriers.
The phase error for a particular test can be calculated by
considering the vectors "a" and "b" representing the digitized
signals measured at points A and B and the vector "c" representing
their differences as the three sides of a triangle, as is shown in
FIG. 6, and by using the geometric cosine law to find the angle y
between vectors "a" and "b". The gain error is the ratio of the
magnitudes of vectors "a" and "b". This is valid only for RMS (root
mean square) values so that a number of samples is required. More
specifically, for each sample, the signal levels (A and B) at
points A and B are measured and squared (A.sup.2 and B.sup.2) by
microcomputer MC. In addition, for each sample, the difference
between the signal levels (A-B) at points A and B is calculated and
squared ((A-B).sup.2). The squared values for all of the samples
are added. The squared magnitude of third side "c" of the triangle
is related to the sum of the squared values of the difference
values (.SIGMA.(A-B).sup.2). The square root of the ratio of the
sums of the squared signal values ((.SIGMA.A.sup.2
/.SIGMA.B.sup.2).sup.1/2) is proportional to the relative gains of
the two channels. The cosine law equation indicated in FIG. 6 is
used to calculate the relative phase angle g from the appropriate
sums of squared values. A relatively large number of samples, for
example, in the order of 500 or more, because the magnitudes of the
signals measured at points A and B are continuously changing,
especially during the DC and asymmetry error -measurements when the
measured signals are responsive to the carriers of the received RF
signal. This method provides sufficient accuracy in an eight-bit
environment.
The use of the cosine law in the above described manner does not
result in reliable results for very small phase difference angles.
The problem is solved in software by means of the addition of a
phase shift to one of the signals measures at points A and B. The
phase shift is latter subtracted from the calculated phase
difference.
Once the gain and phase error values have been measured and stored
in microcomputer MC, the necessary coefficients for the correction
filters can be calculated. FIGS. 7a and 7b show examples of
measured ripple phase and gain errors for one half of the received
spectrum. The phase and gain errors for the other half of the
spectrum are obtained by forming the mirror image of the existing
phase error about the zero point and the mirror image of the
existing gain error about the vertical axis. The resulting full
spectra of the ripple phase and gain errors are shown in FIGS. 8a
and 8b.
The ripple phase and gain error responses have to be combined with
the DC phase and gain error responses. More specifically, the DC
phase error is added to the ripple phase error response shown in
FIG. 8a. The DC gain error is used to multiply the ripple gain
error response shown in FIG. 8b.
If asymmetry errors exists, they can be compensated by using the
asymmetry errors measured by using the picture and sound carriers,
as earlier indicated. More specifically, the asymmetry phase error
for the picture carrier is added to the left side of the ripple
phase error response shown in FIG. 8a and the asymmetry phase error
for the sound carrier is added to the right side. The asymmetry
gain error for the picture carrier is used to multiply the left
side of the ripple gain error response shown in FIG. 8b and the
asymmetry gain error for the sound carrier is used to multiply the
right side. As a result of the combination of the asymmetry errors
with the ripple errors, a step in the middle of the responses may
occur. In many cases the asymmetry errors are small and can be
ignored, and therefore only the response to the picture carrier is
needed. In the present example, it is assumed that the asymmetry
errors are small, and therefor, have been ignored.
The final phase and gain error responses are shown in FIGS. 9a and
9b. The phase and gain error responses shown in FIGS. 9a and 9b
cannot by themselves be used for calculating the error correction
filter coefficients for gain correction unit GC and phase
correction unit PC. Rather, the responses shown in FIGS. 9a and 9b
have to be converted to the phase and gain correction responses
shown in FIGS. 11a and 11b frequency point by frequency point. The
manner in which this is accomplished is illustrated by the vector
diagram shown in FIG. 10. In FIG. 10, the vectors correspond to
signals developed at respectively labeled points of the direct
conversion tuner shown in FIG. 5. With reference to FIG. 10, the
phase (d) and gain errors at each frequency point of the response
shown in FIGS. 9a and 9b are used to calculate a coefficient for
changing the magnitude of vector C and coefficient for changing the
magnitude of a quadrature vector F so that when the resultant
vectors C' and F' are combined, vector G is formed which is equal
in magnitude but opposite in phase to vector D, and therefor
results in cancellation when added to vector D. The desired phase
and gain correction compensation responses shown in FIGS. 11a and
11b correspond to the multiplication factors for vectors C and F
required to produce vectors C' and F' for each sampled frequency on
the frequency axis. The responses shown in FIGS. 11a and 11b are
not symmetrical due to interaction between gain and phase.
The coefficients for phase correction unit PC and gain correction
unit GC can be computed from the responses shown if FIGS. 11a and
11b using the MatLab.TM. FIR2 program. It should be noted that the
coefficients for frequencies below second local oscillator
frequency .omega..sub.N and the ones needed for frequencies above
second local oscillator frequency .omega..sub.N correspond
respectively to the opposite sides of the responses shown in FIGS.
11a and 11b because the inverse positional relationship between the
desired components and the undesired components. This factor is
easily incorporated in the hardware by inverting the signal
developed at point D as is indicated by inverter I as shown in FIG.
5, or by subtracting the signal developed at point D from the
signal developed at point G.
Once the filter coefficients for the selected channel have been
calculated they are stored for retrieval whenever the same channel
is again selected. As a result, the receiver is ready for reception
when a new channel is selected without first calculating "new"
filter coefficients. Temperature affects the drift of components,
such as inductors and capacitors, in the IF filter and therefor
affects the gain and phase error responses. However, it has been
found that it is not necessary to perform the ripple gain and phase
error measurements other than when the receiver is first turned on
or perhaps when a new channel is initially selected. Accordingly,
it is not necessary to interrupt the reception of a program.
Measurement of the DC gain and phase errors is not a problem
because the received television signal can be continuously
monitored without affecting the normal operation of the tuner.
It has been found that the location of the frequency,
.omega..sub.0, of the first local oscillator signals within the
spectrum of the RF signal can be used to optimize the operation of
the direct conversion tuner. For example, the frequency of the
first local oscillator signals should desirably between 1.7 and 2
MHz above the picture carrier (see FIG. 4a) for NTSC television
signals, and between 2 and 2.8 MHz above the picture carrier for
PAL television signals. The frequency of the first local oscillator
signals can be set in accordance with the channel number, for
example, by a phase locked loop tuning control system (sometimes
called a "frequency synthesizer"). In that case, microcomputer MC
may be used to control the frequency of the frequency determining
programmable divider of the phase locked loop. An automatic fine
tuning (AFT) arrangement responsive to the frequency of the picture
carrier may be used to maintain the frequency of the first local
oscillator signals at the desired frequency. The frequency of the
picture carrier may be measured by microcomputer MC during the DC
error measurement operation when the signals developed at points A
and B are sampled.
The direct conversion tuner shown in FIG. 1 has been described so
far with respect to tuning of analog television signals in which
picture, sound and color subcarrier signals are modulated on RF
carriers in accordance with a conventional television standard such
as NTSC, PAL or SECAM. However the direct conversion tuner is also
useful, and may in fact be even more useful, for tuning digital
television signals, such as HDTV (high definition television)
signals. As earlier noted, HDTV systems make full use of the
available channel spectrum, have only a small guard band of a few
hundred kilohertz (KHz) between channels, and require a tuner
response which is flat to the channel edges but which is very steep
at the edges for adequate adjacent channel rejection. A direct
conversion tuner is particularly well suited to such a HDTV
environment because it has a low IF frequency range which allows
the use of simple and effective filters. An IF filter with a sharp
"cutoff" and a large "stop-band" region is much easier to obtain at
low frequencies than at the conventional IF frequencies (38 MHz and
higher).
While the direct conversion tuner described so far is well suited
for tuning digital television signals, certain modifications need
to be incorporated because discrete carriers which can be used to
measure the DC gain and phase errors are usually not transmitted in
a digital television system. However, it has been found that the
spectra of digital television signals, which are typically flat and
have similarities to random noise, may be used to accurately
measure the DC errors. The use of the spectra of the digital
television signals to measure the DC errors may require more
samples to be used than when picture or sound carriers are used.
For example, samples taken over ten to twenty television lines may
be required. In a direct conversion tuner for tuning digital
television signals, the same method of measuring the ripple gain
and phase errors using a plural frequency reference signal
previously described with respect to the direct conversion tuner
for tuning analog television signals is used.
Even if the spectra of the received digital television signals are
not flat for a particular digital television system, the spectra
can still be used to measure the DC gain and phase difference
errors, provided that the shape of the spectra is to modify filter
coefficients.
Digital television signals are more robust than analog television
signals and therefore the generation of unwanted spurious frequency
components is less critical. It is therefore not necessary to
consider the small asymmetric errors and larger tolerances for gain
and phase errors can be accepted.
In a television receiver which is capable of processing both analog
television signals and digital television signals, a single direct
conversion tuner may be used to tune both the analog and digital
television signals. The flow chart shown in FIG. 12 summarizes the
operation of the direct conversion tuner which has been previously
described, and additionally indicates its operation in a dual mode
television receiver. As is indicated in the flow chart, after the
receiver has been energized or a desired channel is selected, the
previously stored gain and phase equalization data for the selected
channel is retrieved from memory and coupled to gain correction
unit GC and phase correction unit PC, the ripple error measurements
are made. Thereafter, the detection of the presence or absence of a
picture carrier causes the selection of either an analog television
signal branch or a digital television signal branch, respectively,
of the program for measuring the DC errors and calculating the
filter coefficients.
While the invention has been described in terms of as specific
embodiment, it is contemplated that modifications will occur to
those skilled in the art. For example while individual gain and
phase correction units are utilized in the embodiment, a single
digital filter may be provided to provide both gain correction and
phase correction. Such a filter may be constructed in either FIR
(finite impulse response) or IIR (infinite impulse response) form.
In addition, while gain correction unit GC and phase correction
unit PC are included in the same channel in the embodiment, one of
the units may be included in one channel and the other may be
included in the other channel. Further, while two analog-to-digital
converter are utilized in the embodiment, it is possible to use a
single ADC which is multiplexed to sample the signals developed at
the measurement points of the two channels. Still further, while
the use of the picture and sound carriers has been described with
respect to the DC and asymmetry error measurements in a direct
conversion tuner for tuning analog television signals, other
components may also be utilized. For example, the color subcarrier
may be used in place of sound carrier in the asymmetry error
measurement. Even still further, while the reference signal is
inserted after the first mixers in the embodiment, the reference
signal can be inserted at other locations such as in the RF stage.
In the same vain, while the measurement points of the embodiment
are located before the second mixers because the second mixers are
implemented in digital form, different measurement points, such as
ones located after the second mixers, may be used. Moreover, while
a direct conversion tuner including the automatic gain and phase
equalization provisions which have been described is particularly
well suited for tuning television signals, it is also useful for
tuning other types of communications signals. These and other
modifications are intended to be within the scope of the following
claims.
* * * * *